Deep Learning Optimizes LEO Satellite Handover

In a recent article published in the journal Aerospace, researchers introduced an innovative handover strategy utilizing deep learning technology to improve the stability and efficiency of communication networks, particularly in power grid scenarios. They aimed to address a major challenge in modern power communication systems: ensuring reliable, high-speed communication in remote and disaster-prone areas.

Study: Deep Learning Optimizes LEO Satellite Handover. Image Credit: NicoElNino/Shutterstock
Study: Deep Learning Optimizes LEO Satellite Handover. Image Credit: NicoElNino/Shutterstock


Low earth orbit (LEO) satellites, operating at altitudes of 500 to 2,000 kilometers, offer advantages such as lower latency and reduced path loss compared to medium earth orbit (MEO) and geosynchronous earth orbit (GEO) satellites. These advantages make LEO satellites ideal for applications requiring high-speed data transmission and low latency.

However, their high mobility necessitates frequent handovers to maintain communication continuity. Traditional handover methods often struggle with the rapid changes in satellite network topology, causing increased latency and potential communication interruptions. To address these challenges, advanced solutions are needed to optimize the handover process in LEO satellite networks.

About the Research

In this paper, the authors presented a robust handover strategy for LEO satellites designed to support communication needs in remote and challenging power grid environments. They developed an innovative approach combining deep reinforcement learning (DRL) and graph neural networks (GNN). This hybrid model aims to optimize satellite handovers dynamically by leveraging the representational capabilities of GNNs and the decision-making strength of DRL.

DRL is integrated with GNN to form a framework that can adapt to and predict the changing topology of LEO satellite networks. Specifically, the researchers used a message-passing neural network (MPNN) to represent the network’s dynamic topology. MPNNs effectively capture interactions between satellites, enabling the prediction of optimal handover decisions. These decisions are then refined using the deep q-network (DQN) algorithm, a well-known DRL method that learns from the environment over time.

Furthermore, the proposed strategy was validated through simulations and experiments, comparing the performance of the MPNN-DQN-based handover strategy with traditional and other DRL-based approaches. They utilized key performance indicators such as handover frequency, communication latency, and network load balancing.

LEO satellite communication-assisted grid scenario figure.

LEO satellite communication-assisted grid scenario figure.

The overall framework of satellite handover decision algorithm.

The overall framework of satellite handover decision algorithm.

Research Findings

The outcomes demonstrated significant advantages of the proposed MPNN-DQN-based handover strategy over traditional methods. This approach effectively minimized the number of handovers, which is crucial for maintaining stable and continuous communication. By accurately predicting the optimal times and targets for handovers, the model reduced unnecessary transitions, thereby enhancing network stability. Frequent handovers can cause service interruptions and increased latency, particularly important for power grid operations.

The proposed strategy also reduced communication delays by optimizing handover decisions and ensuring timely data transmission. The integration of DRL and GNN enabled the model to quickly adapt to changes in network topology, making efficient handover decisions that minimize delays. This capability is especially vital in power grid scenarios where real-time monitoring and control are essential.

Furthermore, the hybrid model improved load balancing across the network. By dynamically adjusting handover decisions based on current network conditions, the MPNN-DQN algorithm evenly distributed communication loads, preventing congestion and enhancing service quality. The model’s ability to learn from and adapt to the changing characteristics of the satellite network environment ensures the communication system remains resilient and reliable, even in highly dynamic and challenging scenarios.


The results of this paper have significant implications for the design and optimization of satellite communication systems, particularly within power grid communications. The enhanced MPNN-DQN-based handover strategy presents several key applications. Firstly, in remote area communication, it ensures continuous and high-quality connectivity where traditional infrastructures are impractical or expensive. This makes it ideal for extending reliable communication to inaccessible areas.

Secondly, during natural disasters disrupting terrestrial networks, the strategy plays a pivotal role in maintaining stable and reliable communication, supporting effective emergency response and recovery efforts crucial for saving lives and coordinating resources. For modern power grids, reliant on real-time monitoring and control, the strategy's improved reliability and reduced latency are essential. It enhances the operation of IoT devices and smart grid technologies, ensuring efficient power distribution and consumption management

Additionally, the strategy supports environmental monitoring by facilitating remote sensing and data collection for ecological research in challenging, remote locations. This capability is crucial for monitoring ecological changes, tracking wildlife, and conducting research in areas that are difficult to access.


In summary, integrating DRL and GNN into the handover management of LEO satellite networks effectively ensured stable and continuous communication for power grid applications. The MPNN-DQN-based strategy outperformed existing methods by reducing handover frequency, lowering latency, and achieving better load balancing, enhancing satellite communication reliability. Moving forward, researchers recommended exploring this novel technology in diverse real-world scenarios and optimizing its performance for different network configurations and traffic patterns.

Journal reference:
  • Yu, H.; Gao, W.; Zhang, K. A Graph Reinforcement Learning-Based Handover Strategy for Low Earth Orbit Satellites under Power Grid Scenarios. Aerospace 2024, 11, 511. DOI: 10.3390/aerospace11070511,

Article Revisions

  • Jul 9 2024 - Inclusion of diagrams from the journal paper.
Muhammad Osama

Written by

Muhammad Osama

Muhammad Osama is a full-time data analytics consultant and freelance technical writer based in Delhi, India. He specializes in transforming complex technical concepts into accessible content. He has a Bachelor of Technology in Mechanical Engineering with specialization in AI & Robotics from Galgotias University, India, and he has extensive experience in technical content writing, data science and analytics, and artificial intelligence.


Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Osama, Muhammad. (2024, July 09). Deep Learning Optimizes LEO Satellite Handover. AZoAi. Retrieved on July 17, 2024 from

  • MLA

    Osama, Muhammad. "Deep Learning Optimizes LEO Satellite Handover". AZoAi. 17 July 2024. <>.

  • Chicago

    Osama, Muhammad. "Deep Learning Optimizes LEO Satellite Handover". AZoAi. (accessed July 17, 2024).

  • Harvard

    Osama, Muhammad. 2024. Deep Learning Optimizes LEO Satellite Handover. AZoAi, viewed 17 July 2024,


The opinions expressed here are the views of the writer and do not necessarily reflect the views and opinions of AZoAi.
Post a new comment

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.

You might also like...
Permafrost Embankment Thermal Prediction via Deep Learning